testing kernel module build in branches/304 and branches/340
luca.boccassi at gmail.com
Wed Mar 30 23:36:55 UTC 2016
On Sun, 2016-03-20 at 23:16 +0100, Andreas Beckmann wrote:
> On 2016-03-20 20:46, Luca Boccassi wrote:
> > I just ran the usual batch, everything looks fine here. amd64 and i386,
> > with kernels 3.2, 3.16, 4.3, 4.4, 4.5-rc7, both -dkms and -source.
> There will be a few more changes before I finally upload that ...
Did another round of builds, all looks good (with 4.5 final also).
Also, I have automated the test builds using cowbuilder via Python and
the pexpect module :-)
In case you are interested, I've pushed it to Github .
It requires at least a pre-configured and pre-installed cowbuilder
chroot, with the kernel headers and other build-dependencies of the dkms
and source packages already pre-installed.
It will spawn in parallel a chroot for each architecture, and run the
test in parallel (per arch) on them. It will test the dkms install
first, and the source build later, for all kernel versions available.
The default options assume that there's a base-sid-$ARCH/result
directory in the pbuilder base path (that's where pbuilder stores built
packages by default), where the built -dkms and -source packages are,
and will copy them to the build chroot with a name of dkms-sid-$ARCH,
and then login in the latter to run the builds. This can all be
overridden via command line arguments in case the packages are already
I've tested it on legacy-YYYxx packages, newer packages where the source
is under -current and older packages (like from the 304 branch) where
it's just under nvidia-$VERSION, seems to work fine!
It depends on python3 and python3-pexpect.
Let me know if you find this useful and if you have any comments!
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 819 bytes
Desc: This is a digitally signed message part
More information about the pkg-nvidia-devel